Reverse engineering the 1988 NeXT keyboard protocol

Article note: Oh, that's fun. The accepted documentation was tellingly nonsense, so they _taught themselves the tools_ to figure it out. That "tracing back to fundamentals" is one of the things I find most delightful about retro-computing. Working all the way back from a glitchy behavior, to a scope trace with unexpected timing, to "This thing is running a very sane serial protocol at the less-sane-sounding 18.958Hz frequency," to correlating to the components and determining that it ties back to a455kHz oscillator with a 24x divider, because 455kHz oscillators were common for AM radios.